146
A N A L Y T I C A L CHEMISTRY
duction coils, the instrument, gives good sharply breaking titration curves for 4he titration of R strong acid by a strocg base. For the titration of a weak acid by a weak base, the reaction of ammonium hydroxide with acetic acid was selected. Approxi-
tion a t both 8 and 40 Mc. (Figure 6). A reversal of slope was obtained, but both end points are satisfactorily sharp. Attainable Precision. In order to study the attainable precision, a titration of hydrochloric acid by 0.1 N sodium hydroxide was interrupted near the end point and carried out with a fine-tip buret with dropwise addition until the end point had been passed. The results shown in Figure 7 indicate a satisfactory precision in the determination of the end point for the majority of chemical purposes. Kumcrous variables which have great significance have not been investigated or discussed in this paper. Among these is the effect of reagent concentrat,ion, limits of detection, and similar physical-chemical variables. Later work may be instituted to investigate these factors, hut to date no information is available. CONCLUSIONS
A titrimeter using a grid-dip type oscillator was constructed and found to have greatcr etability than other instruments tried. When tested on acidimetric and precipitation reactions, the instrument worked satisfactorily and with good precision over a wide frequency range. Additional work covering analytical reactions of interest to these lahoratorieb is contemplated.
I L
I3t l4
10.00.05 10 .I5 .20 .25 .30.35.4 ML. NAOH
Figure 7.
Obtainable Precision
Hydrochloric acid us. sodium hydroxide Theoretical end point 10.22
matelg 0.1 N reagents were used. The titration curve obtained is shown in Figure 5. The titration was carried out a t 8 Mc. Precipitation Reactions. Sodium chloride was titrated with silver nitrate. Satisfactory results were achieved with this reac-
LITERATURE CITED
(1) Anderaon, Kermit, unpublished master’s thesis, Library, Agricultural and Mechanical College of Texas, 1939. (2) Blaedel, W. J., and Malmstadt, H. W., ANAL. CAEM.,22, 734 (1950). (3) Blake, Australian J . Sci., 10,10 (1947). (4) Jensen, F. W., and Parrock, A. L., IND.ENG.CHEM., ANAL. ED., 18,595(1946). (5) Mosesman, M. A., unpublished master’s thesis, Library, Agricultural and Mechanical College of Texas, 1938. (6) West, P. W., Burkhalter, T. S.,and Broussard, L., ANA;. CHEM., 22,469(1950). RECEIVED February 10, 1950. Based on work performed under contract for the United States Air F m s by the NEPA Diviaion, Fairchild Engine and Airplane Corporation, a t Oak Ridge, Tenn.
Determination of Aldehydes in Presence of Ketones Determinatiosi of Acids, Esters, and Alcohols in Presence of Aldehydes JOHN MITCHELL, JR., A N D DONALD MILTON S M I T H Polychemicals Department, E. 1. du Pont de Nemours & Company, Znc., Wilmington, Del.
N
UMEROUS colorimetric techniques have been proposed for the estimation of trace quantities of aldehydes. These include use of the well known Schiff’s, Tollen’s, and Nessler’s reagents. Probably these and other selective aldehyde reagents have found their widest applicability in the determination of formaldehyde (see review by Reynolds and Irwin, 16). The previously reported volumetric procedures usually were based on modifications of Ripper’s bisulfite method (17) or Romijn’s alkaline iodine procedure (19). Although originally proposed for the determination of formaldehyde, these procedures have found limited applications in the analysis of other carbonyl compounds. The bisulfite method, reviewed in several papers (8, 10, 16, &Ihas ), been w e d in the determination of many aldehydes and methyl ketones. An interesting variation in technique was suggested by Siggia and Maxcy (80) for the determination of aldehydes. These investigators employed a sulfite-sulfuric acid reagent and determined e x c w acid potentiometrically. Fairly sharp inflection points were observed on titration of solutions containing the aldehyde-bisulfite addition compound. With the exception of cyclohexanone, which a p
peared to react quantitatively, other ketones studied did not interfere seriously in the titration, provided their concentrations were below 10 mole % of that of the aldehyde. The Romijn iodine oxidation procedure (19) has found restricted application in the determination of other aldehydesfor example, Korenman (9) and Rogers (28)applied this reagent to the determination of acrolein and furfural. Another reported method of limited applicability for the determination of aldehydes waa based on the Cannizzaro reaction; by employing rather drastic conditions, Palfray and co-workers (14) were able to determine many aromatic and a few aliphatic aldehydes. The familiar Tollen’s reagent has been employed widely in a nearly specific qualitative test for aldehydes. I n this determination silver oxide, a mild oxidizing agent, is reduced by the aldehyde to free silver, RCHO
+ A e O +RCOOH + 2Ag
(1)
One of the few reported quantitative applications of this reaction was described by Alekse’ev and Zvyagina ( 1 ) for the
V O L U M E 22, NO. 6, J U N E 1 9 5 0 The silver oxide oxidation of aldehydes has been made the basis for a quantitative acidimetric macromethod for their determination. The resulting acid is titrated with standard sodium hydroxide. With the exception of cyclohexanone, which appears to react slightly, no interference from ketones has been observed. Acids originally present in the sample consume an equivalent quantity of alkali. Some of the lower esters interfere through their
determination of 5- to 40-mg. quantities of butyraldehyde. The resulting butyric acid was oxidized t o acetone with hydrogen peroxide which, in turn, was determined by an alkaline iodine (f2), hydroxylamine hydrochloride (8, I f ), or furfural ( 4 ) procedure. According t o Equation 1, one mole of acid is formed for each mole of aldehyde group oxidized. This reaction has been made the basis of a new general acidimetric macromethod for the determination of aldehydes: The sample, containing aldehyde, is treated with an excess of silver oxide at 60" C. After oxidation is complete, the organic anions are converted to salts by the measured addition of an excess of standard aqueous sodium hydroxide. After the insoluble silver oside has been removed by filtration, the excess sodium hydroxide is determined by titration with standard hydrochloric acid. REAGENTS
Silver oxide, c.P., may be obtained from Eimer and AmendFisher Scientific Co., 635 Greenwich St., New York, N. Y. Standard 0.5 ,Ir aqueous sodium hydroxide, standard 0.1 N aqueous hydrochloric acid, and phenolphthalein indicator solution are prepared in the usual way. Dioxane is purified by the method of Eigenberger (6, IS). PROCEDURE FOR DETERMINATION O F ALDEHYDES
A weighed sample, containing up to 5 millimoles of aldehyde, is transferred to a 250-ml. glass-stoppered volumetric flask containing 3 grams of silver oxide and exactly 50.0 ml. of distilled water measured a t room temperature (25.0 ml. of water plus 25.0 ml. of dioxane for water-immiscible samples). The firmly stoppered flask, together with a blank containing silver oxide and 50.0 ml. of solvent, is placed in a water bath at 60" C. The mixtures are shaken vigorously every 5 to 10 minutes. After an hour, the flasks are removed and allowed to cool to room temTable I.
Analytical Data for Aldehydes FTW
Caibiyl Found, Wt. % .4nalysis by Other Aldehyde, Method, .4g10 Free NO. Compound Wt. yo method acidity Total 1 Chloral 96.Ba ( 4 ) s 9 6 . 3 1 0 . 3 ... 9 9 . 3 Q (4) 9 9 , 5 t 0 . 3 0 . 5 2 Propionaldehyde 100:0 9 8 . 7 a (4) 9 8 . 8 1 0 . 3 0 . 8 3 n-Butyraldehyde 99,6 4 n-Butyraldehydec (8) 100.1 1 0 . 2 0 . 0 100.1 5 Isobut raldehyde 9 6 . 2 " (3) 9 6 . 4 1 0 . 2 0 . 9 97.3 6 a-MetXyl acroleind 7 9 . 5 " (2) 8 0 . 3 1 0 . 3 0 . 2 80.5 7 Aldold (4) 102.1 - 0 . 3 0.0 102.1 98 7a (4) 9 8 . 6 1 0 . 2 8 Crotonaldehyde 1.4 100.0 9 Isovaleraldehyde (2) 9 6 . 1 1 0 . 2 3 . 8 99.9 8 9 . 2 c (8) 8 9 . 7 1 0 . 4 6.6 10 2-Ethyl hexanald 96.3 1 1 Dextrose h drated (2) 9 8 . 9 1 0 . 5 0.0 98.9 70 7 ' 12 CitronellalJ (2) 7 0 . 9 - 0 . 0 ... 13 Benzaldehyde 8 9 . 8 " (12) 9 0 . 0 * 0 . 5 10.1 ioo:i 14 BenzaldehydeC (4) 9 9 . 6 1 0 . 3 0 . 3 3 99.9 15 p-Chlorobenzaldehyded (4) 8 7 . 7 1 0 . 1 1 1 . 7 99.4 9 4 . 4 c (2) 9 4 . 5 1 0 . 4 2 . 5 16 p-Nitrobenzaldehyded 97.0 17 Salicylaldehyde 9 8 . 4 " (4) 9 8 , 2 1 0 . 3 0 . 5 98.7 18 Tolualdehyde (2) 8 9 . 1 -0.2 1 0 . 1 99.2 4 8 . 7 * (2) 4 9 . 0 1 0 . 0 19 Phenyl acetaldeh ded ... ... 20 Cinnamaldehydez (2) 6 1 . 4 1 0 . 0 3 3 . 3 94.7 91.2O (2) 9 0 . 8 - 0 . 2 21 Furfural 92.1 1.3 'I By hydroxylamine hydrochloride-pyridine method ( d ) , 0.5 hour at room temiierature. h Figures in parentheses represent number of individual determinations. C Freshly distilled. d Practical grade. .4ldeh dea not marked were c P. grade. lo;oF3~. hydroxylamine hy&ochloride-pyridine mdthod ( d ) , 2 hours at
747
saponification by the silver reagent. Methods are presented also for the determination of other compounds in the presence of aldehydes. Acids are titrated with standard sodium methylate solution. Esters are saponified after active carbonyl compounds have been converted to the oximes. Alcohols are determined after the aldehydes have been separated by treatment with silver oxide or with sodium bisulfite.
perature. Exactly 25 ml. of 0.5 N sodium hydroxide are added plus sufficient distilled water or dioxane t o make 100.0 ml. of total liquid. The mixture is shaken and filtered. Accurately measured 25.Gml. portions of the clear filtrate are titrated with standard 0.1 N hydrochloric acid t o the phenolphthalein end point. There is a l.Oy0 contraction on mixing equal volumes of dioxane and water, with proportionately smaller contractions for intermediate mixtures. The volume of filtrate taken for titration should be corrected appropriately for the change in volume due to dioxane. As an alternative procedure, the total liquid is recovered by filtration and the filtrate is titrated with standard 0.2 N hydrochloric acid. In this case careful measurement of the distilled water or dioxane is unnecessary. The net decreese in sodium hydroxide (blank minus sample titer) is a measure of the moles of aldehyde in the sample. ANALYTICAL RESULTS
The new procedure has been used successfully for the analysis of a wide variety of aldehydes. Results on representative compounds are given in Table I. In most cases C . P . chemicals were analyzed and compared with data obtained by a hydroxylamine hydrochloride procedure. The free acid found in these samples also is recorded. n-Butyraldehyde (No. 4, Table I ) had been carefully distilled through a 36 X 0.5 inch Fenske ring-packed column which was blanketed with nitrogen. When analyzed shortly after fractionation, the heart cut of the distillate was acid-free and gave results for n-butyraldehyde averaging 100.1%. Benzaldehyde (No. 14, Table I ) had been distilled in the same manner; however, a small quantity of acid was present in this distillate. When the per cent acid was added t o the average per cent aldehyde found, the total w w 99.9%. A few aldehydes, other than those listed in Table I, were tested and found t o give erratic results. A sample of 37% aqueous formaldehyde analyzed 33.6 * 0.5% in four determinations. Acetaldehyde could not be determined by the general technique, probably because of its volatility. Values of 88 * 1% were obtained after 1 hour a t 60" C. on freshly distilled acetaldehyde, analyzing 99.8 * 0.2% by the bromophenol blue method ( 2 ) . Similar results were observed on samples shaken with the silver oxide for 1 hour at room temperature and then heated for 0.5 hour a t 60" C. However, eamples shaken continuously with reagent for 2 hours a t room temperature gave recoveries of 99.0 * 0.4%. Probably a slightly longer time a t room temperature would be sufficient for complete reaction. Technical grade citral analyzed 75% compared to 92.5% by the hydroxylamine hydrochloride-pyridine method ( 2 ) . Results on citronellal (No. 12, Table I), on the other hand, compared favorably with those of an independent method. C . P . quality anisaldehyde, vanillin, and dimethylbenzaldehyde gave values of 65, 120, and 125'4, respectively. Where the total free carbonyl of those compounds listed in Table I was also determined by an independent method, the results by the new method for aldehydes checked the other results to 0.270 in nearly all cases. The data recorded in Table I indicate that the new method is applicable to aldehydes in general with an average precision and accuracy of about *0.30/,. Several known propionaldehydeacetone mixtures were prepared in dioxane solution. Suitable duplicate aliquots were
ANALYTICAL CHEMISTRY
748 Table 11. Determination of Acids in Presence of Aldehydes 8ubotan ce Formaldehyde Butyraldehyde Butyraldehyde Iaovaleraldehyde Bensaldehyde Benxaldehyde Tolualdehyde imethylbensaldehyde a Added acid WM acetic. b Freshly distilled in an atmwphere Freshly distilled.
NO.
E;.
0.28 * 0.00 3.53 a0.01 0.00 h O . 0 0 9.82 a O . 0 1 10.83 f 0.02 0.33 a 0 . 0 0 13.98 r 0 . 0 2 0.74 a O . 0 0
34.1 28.6 18.2 35.9 38.3 16.4 29.7 35.8
34.1 28.6 16.2 35.8 38.3 18.5 29.6 35.9
of nitrogen.
analyzed by the standard procedure. The following are typical results : Acetone, Wt. % 96.5 88.6 55.0
33.3 11.5
Propionddehyde, Wt. 75 Added Found 3.5 3.6 r 0 . 2 11.4 11.2 -0.2 45.0 45.0 -0.1 68.7 87.0 I0.3 88.5 88.3 -0.2
DETERMINATION OF ACIDS, ESTERS, AND ALCOHOLS IN PRESENCE OF ALDEHYDES
Interfering Substances. The new method is subject to Interference from peroxides, sodium methylate, acids, and esters. No interferences were observed from alcohols, ethers, and most ketones. Acetone, methyl ethyl ketone, diethyl ketone, and acetophenone did not react. However, cyclohexanone appeared to react slightly. Quadruplicate analyses by the aldehyde method of peroxidefree cyclohexanone gave values of 1.5 * 0.2%. Peroxides in general probably interfere-for example, cyclohexanone containing 1% peroxide (calculated as H ~ O Z ) gave apparent aldehyde values aa high as 25%. Sodium methylate in methanol solution reacted with the silver oxide reagent; a lustrous silver mirror was plated on the g1w container. The observed behavior suggests the following reaction: NsOCHl
+ 2Ag20 +4Ag + HCOONa + HzO
(2)
Acids interfere directly in the aldehyde procedure by consuming an equivalent quantity of alkali. Therefore, the sodium hydroxide titer must be corrected for the quantitative acid reaction. Acids cannot be determined accurately in the presence of active aldehydes by titration with aqueous sodium hydroxide. Several aldehydes were titrated with aqueous sodium hydroxide. No stable phenolphthalein end points could be obtained and in general the results were 0.3 to 0.6% high. However, when acid-aldehyde mixtures are titrated with standard sodium methylate in dry methanol, sharp phenolphthalein end points are obtained which are a quantitative measure of the total acidity of the sample. Duplicate 4- to 10-gram portions of the aldehydes listed in Table I1 were titrated with 0.5 N sodium methylate to determine their initial acid contents. Then larger known amounts of acid were added to separate portions of the aldehydes and the solutions titrated using the acid titer of the initial aldehyde as the blank.
The more easily hydrolyzed esters are saponified by the silver oxide reagent of the procedure for aldehydes. For example, ethyl acetate, methyl propionate, methyl valerate, and cyclohexyl acetate reacted quantitatively in either aqueous or waterdioxane solution (see also Table IV). Therefore, the net alkali consumed in the aldehyde determination must be corrected for the ester equivalent. This ester equivalent can be determined accurately only for esters which are quantitatively saponified by the silver oxide reagent. Direct saponification methods cannot be used for the determination of esters in the presence of aldehydes. Aldehydes seriously interfere by absorbing alkali in amounts which usually bear no stoichiometric relation to the concentrations of aldehydes. The behavior of aldehydes toward heating with sodium hydroxide is illustrated by the following duplicate analyses of several aldehydes, which contained less than 0.5% of ester. The Bryant-Smith saponification technique ( 9 ) was used which involves heating the sample a t 60" C. for 30 minutes with 2 N sodium hydroxide in 90% methanol. Aldehyde ( 2 9 9 5 Wt. %) Propionaldehyde bobutyraldehyde Benraldehyde Dimethylbenzaldehyde
Apparent Ester (as W t . % ' CS Ester) 21.0 a 0 . 2 23 8 -0.3 18.0 0.5 1.0 - 0 . 1
However, Smith and Bryant (22) found that the oximes of aldehydes were sufficiently stable toward alkali to permit esters to be saponified without interference. Although ketones did not interfere in the saponification, i t proved easier not to attempt to distinguish between aldehydes and the more active ketones where both were present, but to convert both to the corresponding oximes, using aqueous hydroxylamine as the reagent. It waa necessary to avoid e x c w hydroxylamine, because the free base in the presence of strong alkalies reacts with eaters to form acyl hydroxamic acids, HzNOH
+ RCOOR' +RC(OH)=NOH + R'OH
(3)
The hydroxamic acids are too weak to be titrated successfully. When permitted to form, they introduce a considerable error into the ester determination. The correct amount of hydroxylamine to be used is predetermined by the hydroxylaminepyridine method (8).
Determination of Esters in Presence of Aldehydes. REAQENTS. Hydroxylamine hydrochloride reagent, 2 N , is prepared from 138 grams of C.P. salt in sufficient distilled water bo make 1 liter of solution. The exact normality of the 2 N hydroxylamine hydrochloride is determined by titrating the free acidity in a 20.0ml. portion of the solution to the bromophenol blue end point with aqueous 0.5 N sodium hydroxide, then adding an excess of C.P. acetone (6 to 10 d.) and , allowing the mixture to stand 5 minutes. Finally the solution is again titrated to the broniohenol blue end point with standard aqueous 0.5 N sodium ydroxide The latter alkali titer determines the normality of the reagent solution, one mole of hydroxylamine hydrochloride being equivalent to a mole of sodium hydroxide. As an alternative method for determining the hydroxylamine reagent normality, a second 20-ml. portion may be titrated to the phenolphthalein end point. The net titer between the broniophenol blue and phenolphthalein end points also is a direct measure of h droxylamine hydrochloride. Thymolphtialein indicator is a 1%solution of the C.P. material in absolute ethanol. Ap roximately 2 N sodium hydroxide reagent is prepared by dissohng 80 grams of C.P. pellets in sufficient 90% methanol10% water to make 1 liter of solution. Determination of Acids in Presence of Aldehydes. REAGENTS. PROCEDURE. A carefully measured volume of sample is Approximately 0.5 N sodium methylate is prepared b dissolving weighed into a 250.0-ml. volumetric flask or Pyrex brand pres27 grams of the C.P. powder (Mathieson Alkali dorka) in 1 sure bottle. The amount of active carbonyl is determined by means of the hydroxylamine-pyridine method (X), allowing a liter of dry methanol solution. The reagent is standardized against Bureau of Standards benzoic acid or standard 0.5 N 5-minute reaction time a t room temperature. aqueous hydrochloric acid. Then the calculated e uivalent amount of 2 N aqueous hydroxylamine hydrochlori2e solution is delivered accurately into Phenolphthalein indicator solution. The sample, containing up to 20 millimoles of an empty 250.0-ml. glass-stoppered volumetric flask by means of PROCEDURE. a 10.0-ml. microburet. Thymolphthalein is added and the soluacid, is transferred to an Erlenmeyer flask and titrated with tion is carefully neutralized to a blue color with 2 N sodium standard 0.5 N sodium methylate to the phenolphthalein end hydroxide in 90% methanol (10% water). The solution is next point.
.
149
V O L U M E 2 2 , NO. 6, J U N E 1 9 5 0 back-titrated with a few drops of 0.5 N aqueous hydrochloric acid until the blue color just disappears. The hydroxylamine is then substantially all in the free form. The sample to be analyzed is now added. For optimum precision and accuracy, this sample should contain about 10 milliequivalenta of acid plus ester. Th? sample also contains aldehyde plus ketone exactly e bivalent to the hydroxylamine liberated in the above step. ?%e mixture is allowed to stand for 5 minutes at room temperature. Then 20.0 ml. of 2 N sodium h droxide are added, and the flask is stoppered and heated for 0.5 gour in a water bath at 60" C. At the end of this time, the flask is removed, cooled, and titrated with standard 0.5 N hydrochloric acid. [These last steps are essentially identical with the reviously published general saponification procedure (3).] T\e thymolphthalein color change is sharper in the presence of oximes than the phenolphthalein end point. At least one blank, containihg 20.0 ml. of the 2 N alkali, is run with each set of samples. No hydroxylamine should be employed in the blank, because the f;ee b&e under oes decomposition when heated in strong sodium hydroxide sofutions. The reagents employed prior to the saponification step are used in concentrated form t.0 avoid excessive dilution. In general, the final caustic normality for saponification should be no less than 1 N . A number of synthetic mixtures were analyzed successfully by the above procedure. Typical results are given in Tables I11 and IV. hIost of the samples of Table I11 were prepared volumetrically and then calculated to a weight basis by means of densities. Nos. 3, 4, 10, and 11 were prepared on a gravimetric basis. The data of Table I11 have an average reproducibility of about &0.3% in weight composition and seldom differ from the true composition by more than 1%. The largest errors are asociated with the volumetric samples containing relatively small concentrations of ester and probably represent, in part, errors in volume measurement. A series of known mixtures of isovaleraldehyde and methyl valerate was prepared and analyzed by both the silver oxide method (aldehyde plus ester) and the modified saponification procedure. Two 100-ml. dioxane solutions were prepared. One
T a b l e 111. D e t e r m i n a t i o n of Esters in Presence of Aldehydes Mixture aa Prepared KO.
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Aldehyde, wt. % Propionaldehyde Propionaldehyde n-Butyraldehyde n-Butyraldehyde Isobutyraldehyde Isobutyraldehyde Isobutyraldehyde Citral BensaIdehs.de Benzaldehyde Benzaldehyde Benzaldehyde Dimethylbenzaldehyde Dimethylbenzaldehyde
T a b l e IV. Aldehyde 95.7 89.0 87.8
54.7
53.5 38.3 13.9 9.1
4.6 0.0
Eater b Analysis,
Ester. wt. % 95.0 43.5 95.0 54.5 94.9 78.8 43.1 87.4 96.0 94.0 65.0 38.0 95.9 37.2
d.%
5.0; 56.5 5.0d 45.5; 5.1 21.2" 56.gC 12.6: 4.0 6.0' 36.0' 62.0: 4.1 62.8*
(3)b 6 . 7 (3) 5 6 . 8 5.4 45.2 6.2 22.0 (3) 5 7 . 6 (3) 1 3 . 7 (3) 3.3 6.2 34.8 3) 6 1 . 3 {3) 3.7 (3) 6 1 . 9
{!
$\
tO.2 t0.3 dO.2 *0.3 *0.5 t0.3 h0.3 t0.4 ==0.2 *0.3 t0.3 t0.3 t0.4 tO.2
Analysis of IsovaleraldehydeMethyI Valerate Mixtures
Added, Weight % Eater 0.0 7.0 8.2 44.1 45.4 61.1 85.8 90.6 95.2 99.9
Acid 4.3b 4.0 3.9 1.2
1.1
0.6 0.3 0.3 0.2 0.1b
Found, Weight % Aldehyde' Ester 95.8 88.8 88.2 54.5 53.5 37.8 13.5 8.7 4.2
-0.2
0.3 5.9 8.6 45.0 46.3 61 .O 85.5 90.8 95.1 99.9
Corrected for known ester added. Determined by titration of concentrated sample; all other acid valuea were calculated.
Table V. D e t e r m i n a t i o n of Alcohols in Presence of Aldehydes Aldehyde, Weight % Propionaldehyde 3 0 . 0
Methanol
73.5
Ethanol Propanol Butanol
Propionaldehyde 6 4 . 3 Butyraldehyde 4 7 . 5 Butyraldehyde 2 4 . 7 Butyraldehyde Valeraldehyde 3.5 5-Trimethyliexanal
51.2 78.3 15.0 73.6 45.7
9.6 28.5 58.7 97.3
Alcohol, Weight % Found Added
Method for Aldehyde Removal
NaHSOI
Pentenol
NaHSOI
2-Eth lhexanol 3,5,5-6rimethylhexanol.
NaHSOi
70.0 26.5 35.7 52.6 75.3 48.8 21.7 85.0 26.4 54.3
68.9 26.0 35.0 51.7 74.8 48.4 21.3 84.6 26.1 54.0
90.4 71.5 41.3 2.7
90.3 71.3 41.2 2.4
contained 8.92 grams of isovaleraldehyde and the other 14.0 g r a m of methyl valerate. Aliquants of each were mixed and analyzed. Results are given in Table IV. The ester-free aldehyde was analyzed by the 2.5 p H hydroxylamine hydrochloride method ($3). I n general, the accuracies of the found values for both aldehyde and ester compare favorably with those data of Table I and 111. Aldehydes interfere in most acylation procedures for alcohols. [A notable exception is the phthalic anhydride method of Elving and Warshowsky ( B ) , which is applicable to many primary and secondary alcohols. ] The silver oxide method frequently may be employed conveniently for the removal of aldehydes from samples prior to the determination of free alcohols, except in the presence of easily hydrolyzed esters.
Determination of Alcohols in Presence of Aldehydes. T h e sample, containin up to 10 millie uivalents of aldehyde but no more than 20 mifiequivalents of 814, alcohol, is weighed into a distillation flask containing about a 100% excess of silver oxide (4.6 grams of Ag,O for 10 millimoles of aldehyde). Fifty milliliters of water are added and the mixture is heated for 0.5 hour a t 60" C. At the end of this time, the sample flask is removed and connected to a small still. If the alcohol is methanol or ethanol, 50 ml. of C.P. benzene are added and the mixture is distilled carefully until 5 to 10.0 ml. of benzene-water azeotrope have been removed. Azeotropic Data (7) Benzene-methanol B P 58.34O C 39,6'7 CHIOH Benzene-ethanol B:P: 68.24' C:: 32.403 C~HIOH Benzene-water B.P. 69.25O c.,S . S % ~ H ~ O
The distillate is diluted to exactly 50.0 ml. with purified dioxane and suitable ali uota are used for the subsequent hydroxyl determination. If txe alcohol is propanol or butanol, the aqueous silver oxide-treated sample mixture is distilled directly until 25.0 ml. of distillate have been obtained. Then 25.0 ml. of C.P. benzene are added to the distillate plus sufficient anhydrous potassium carbonate to saturate the water layer. The benzene layer, which now contains essentially all of the alcohol, is removed and used in the analysis. Alcohols containin five carbon atoms or more are only slightly soluble in water ancfare best se arated by benzene extraction of a bisulfite-treated aqueous sorution. The sample is shaken vigorously with sufficient 10% aqueous sodium bisulfite solution to react with all of the aldehyde. Usually a 50% molar exceas of bisulfite solution is adequate. Then the mixture is extracted with 25-ml. portions of benzene. The hydrocarbon extracts which now contain all of the alcohol, are combined and measured portions are used for the alcohol analysis. Typical results on several aldehyde-alcohol mixtures are given in Table V. In all cases the acetyl pyridinium chloride method (41)was used for the determination of the alcohol. DISCUSSION
The new procedures have been used succeasfully for the analysis of complex mixtures involving aldehydes and for the removal of aldehydes to prevent their interference in analysea for other functional groups.
ANALYTICAL CHEMISTRY
750
LITERATURE CITED
(3) Ibid., 58, 1014-17 (1936). (4) Chelintzev, V. V., and Nikitin, E. K., J . Gen. Chem. (U.S.S.R.), 3, 319-28 (1933). (5) Eigenberger, E., J . prakt. Chem., 130 (2), 75-8 (1931). (6) Elving, P. J., and Warshowsky, B . , ANAL.CHEM.,19, 1006-10 (1947). (7) Horsley, L. H., Ibid., 19, 508-96 (1947). (8) Joslyn, M. A., and Comar, C. L., IND.ENG.CHEM., ASAL.ED., 10, 364-6 (1938). (9) Korenman, I. M., J . Applied Chem. (U.S.S.R.), 8, 1470-7 (1935). (10) Lasaari, G., Ann. chim. applicata, 32, 78-86 (1942). (11) Marasco, M., Ind. Eng. Chem., 18, 701-2 (1926). (12) Messinger, J., Ber., 21, 3366-72 (1888). (13) Mitchell, J., Jr., and Smith, D. M., “Aquametry,” p. 153, New York, Interscience Publishers, 1948. (14) Palfray, L., Sabetay, S., and Sontag, G., Chimie et Industrie, 29, 1037-8 (1933). (15) Parkinson, A., and Wagner, E., IND.ENG.CHEM.,ANAL. ED., 6, 433-6 (1934). (16) Reynolds, J. G., and Irwin, M., Chemistry & Industry, 1948, NO. 27, 419-24. (17) Ripper, M., Monatsh., 21, 1079-85 (1900). (18) Rogers, H. R., IND.ENG.CHEM.,ANAL.ED.,16, 319-21 (1944). (19) Romijn, G., 2.anal. Chem., 36, 18-24 (1897). (20) Siggia, S., and Maxcy, W., ANAL. CHEM.,19, 1023-4 (1947). (21) Smith, D. M., and Bryant, W. M. D., J . Am. Chem. Soc., 57, 61-5 (1935). (22) Smith, D. M., and Bryant, W. M. D., unpublished method. (23) Smith, D. M., and Mitchell, J., Jr., ANAL. CHEM.,22, 750 (1950).
(1) Alekse’ev, S. V., and Zvyagina, S. I., Sintet. Kauchuk, 5, 1926 (1936). (2) Bryant, W. M. D., and Smith, D. M., J . Am. Chem. Soc., 57, 57-61 (1935).
RECEIVED October 20, 1949. Presented before the Division of Analytical and Micro Chemistry at the 116th Meeting of the AMERICAN CHEMICAL SOCIETY, Atlantic City, N. J.
A synthetic mixture containing 50% methyl acetate, 12% methyl formate, 28% acetone,- and 10% propionaldehyde analyzed 9.8 * 0.4% aldehyde after correction for the ester interference. Another mixture containing 60.0% acetic acid, 12.0% methyl formate, 24.0% methyl acetate, 3.0% acetone, a n d 1.0% propionaldehyde was first neutralized with sodium methylate and then analyzed. Values of 1.0 * 0.5% propionaldehyde were found, after correction for the saponification of the esters. The silver oxide procedure also was employed t o permit the determination of acetone in the above complex mixtures. Portions of both samples were treated with silver oxide according t o the standard method. Then the acetone was separated by distillation through a simple column packed with Pyrex brand glass rings, sufficient water being taken over to assure complete removal of acetone. Analyses of the distillate by the 2.5 p H hydroxylamine hydrochloride method (29) gave values of 27.3 * 0.3% (added 28.0%) and 3.0 * 0.2% (added 3.0%) acetone. ACKNOWLEDGMENT
The authors are grateful t o the following for their aid in carrying out the experimental program: C. E. Ashby, W. L. Autman, Alberta Barta, N. D. Boyer, Walter Hawkins, and A. N. Oemler.
Determination of Carbonyl Compounds in the Presence of Organic Acids DONALD MILTON SMITH AND JOHN MITCHELL, JR. PolychemiculsDepartment, E. I . du Pont de Nemours & Company, Znc., Wilmington, Del. A modified acidimetric oximation procedure is presented, in which the pH of the reagent is reduced to a point where carboxylic acids barely interferee.g., only 0.05 ml. of 0.5 N sodium hydroxide is consumed per gram of most carboxylic acids. The hydroxylamine hydrochloride reagent is prepared in 80% ethanol solution, thymol blue indicator is added, and the apparent pH is adjusted to 2.50 by the addition of aqueous hydrochloric acid. Most aldehydes and methyl and ethyl ketones react quantitatively at mom temperature. The ketoximes are sufficiently basic at pH 2.50 to interfere slightly and,
H
YDROXYLAMINE hydrochloride is a reliable and widely applicable reagent for the determination of aldehydes and ketones. I n the equations RCHO
+ H2JOH.HCI+RC(H)=KOH
+ Hz0 + HCI
(1)
+ HnNOH.HCI+RR’C=NOH
’+ HsO + HCI
(2)
and RCOR’
one mole each of acid and water is released for each equivalent of carbonyl compound reacted. Originally this reagent was used by Brochet and Cambier ( 1 ) for :he quantitative determination of formaldehyde. Later investigators demonstrated that hydroxylamine hydrochloride (or sulfate) can be the basis of valuable procedures for the determination of all types of carbonyl compounds. These procedures, summarized by Mitchell
therefore, to require a small correction factor. This system also may be employed to determine certain acetals and ketals. In many cases the hydrogen ion concentration of the reagent is sufficiently high to effect nearly quantitative hydrolysis of these combined carbonyl compounds in 1 to 2 hours at 98” to 100’ C. Thus a scheme is available for determining in a single sample both free and combined carbonyl compounds. After titration for the free carbonyl groups, the sample plus reagent is heated on the steam bath, cooled, and titrated again. The second titer is a direct measure of acetals or ketals. and Smith (7), almost invariably involved titration of the acid freed during oxime formation according to Reactions 1 and 2. I n one case the water of reaction was determined by titration with Karl Fischer reagent (8). For trace quantities of acetone, differential p H measurements have been employed (9, 5 ) . All the acidimetric methods previously reported were subject to interference from acidic and basic components of the samples to be analyzed. However Eitel’s ( 4 ) potentiometric titration method reportedly could tolerate small quantities of acids. For a nearly completely aqueous system, this investigator assigned a p H of 4.1 as the equivalence point for free hydrochloric acid in the presence of hydroxylamine hydrdchloride. In carrying out an analysis Eitel ( 4 ) added aqueous hydroxylamine hydrochloride (or sulfate) to an alcoholic solution of the sample. Depending on the type of carbonyl compound, the mixture was